Mosfet/PLL/ other heatsinks?

I just ordered my IC7-g MaxII today. I am looking to get 4+Ghz. I know I will be needing some mosfet sinks, But is there really a need for a PLL heatsink and a southbridge heatsink?

http://www.sidewindercomputers.com/memandvidcar.html

yes
The PLL chip regulates power going to CPU. It might make the difference in an OC and a stable OC.

I have done the heatsink thing to a couple of IC7 mobos. The mosfets around the CPU socket and ATX power connector are the most important. If you have extra sinks, then there is one or two near the AGP slot that get hot too. At high FSB, cooling the PLL chip (clock generator) helps too. But, I have not seen any benefit in sinking the southbridge other than it looks cool. However, improving northbridge cooling does help at high FSB.

Are the benefits of using AS5 + epoxy vs thermal pads significant and worth the hassel? If so, what kind of mix do I need to use? Thanks

If you Arctic Silver/Alumina thermal epoxy, then that's all you need. If you have regular epoxy or super glue, then you put thermal paste in the center and epoxy or super glue on each corner. That's far better than thermal tape.

My baby came in today... I am thinking of getting either the Swiftech MCX-159CU for the NB which I can use on this 875 or for my future AMD x2 upgrade or Swiftech MCW20 that I can only use on this board. I already have watercooling and they both would come out to be the same price.

I know whats the practical choice, but would the waterblock be that much better for reaching 280 FSB+?
Thanks

Back in my 2.4C M0 stepping days (actually still have it too), I was pushing 290 to 300 FSB on an IC7 with improved or modded N/B air cooling as were several others on the forum. With the heat of those Prescotts, I'm not sure I would watercool the N/B. Leave all that cooling capacity for the CPU.

I actually have every last one of my fets sinked (not sure if it made any difference on the ones not around the cpu). Those microsinks aren't small enough to get them all sinked. I had to make custom ones from an old Dell heatsink I had around the house. It's times like these that make me wish I had a bandsaw with a metal cutting blade! I wish you could find one of these for under 75, that would be tops.

Right out of the box and this is what I posted... I haven't done anything...
I love Abit. If I didn't have a test tomorrow, I would find the absolute max.

Still need to get some sinks for the mosfets, PLL, etc.. and the NB. I can't wait.

Nice, dude. Just noticed what you're cooling with.

I used Alumina Epoxy because I was a weenie about it, go for the Arctic Silver Thermal Adhesive if you don't have a shakey hand. It should deliver a degree or two better cooling.

Just wanted to share my slow progress. I wanted to go for the gold on the first try and to boot 4.5Ghz @ 1.6v into windows. That didn't work, so I scaled down to 4Ghz @ 1.6v and noticed that my PWM temperature was ridiculous. They were 62C!!! and climbing on load, which isn't acceptable. Note, I haven't sinked my mosfets or anything yet which is probably why they were at 60+C. I can't afford to blow another mobo nor can I afford some sinks right now (kinda sad, but thats the college life). So I eventually scaled down to the sweet spot for 4Ghz which is now stable 1.48v. My PWM load temperatures were still high, at 52C. I wanted to keep my mosfets below 50C which seem to be the target high based on multiple articles on this forum ( I love Ocforums!). Maybe a fan to cool the mosfets? ... then I thought, I could do one better, a fan to cool the mosfets, the NB, and Ram. I dug around my boxes of c*** and found a 120mm fan and zalman fan adaptor(not sure what they are called) but never the less.

CPU full load
Before Fan:
PWM temp: 52C
AFter Fan:
PWM temp: 48C

Those numbers speak for itself, if the mosfets are going to drop 4C without even being sinked. Imagine what the temperature with sinks and a fan.

When I get enough money, I will sink my mosfets and no longer be hampered by my PWM temperatures and reach my max OC!

Nice results so far, 4 gig from a 3.0E is excellent. But, I wouldn't go much above 1.5v long term even with the mosfets sinked.

How much above 1.5v would be safe for long term use? Which makes sense because the general rule is +- 10% which would be 1.51v but at the same time. I've seen some posts where people would keep their core at 1.6v.

What would need to be modified in order to scale up to 1.6v or even 1.65v safely given I am able to cool everything down.

I've always thought temperature was my largest and only enemy given I am not too crazy with the voltages since every circuit has a limit.

Is this motherboard's circuit layout simply unable to sustain the higher current from higher voltage? I should hope so because the BIOS allows voltages way higher than 1.6v.

There is much I need to learn.
Thanks for the help

Simply, the IC7 series socket 478 motherboard was designed for the Northwood processor. That's one reason the vcore goes up pretty high is because Northwoods had a higher default vcore to start with. The power hog Prescott was not even a twinkle in Intel's eye yet. Abit and other makers were able to add Prescott support to the i865/i875 chipsets with a BIOS update. This works fine at default voltages. But, the power circuitry is still not up to the challenge of hardcore overclockers willing to push high power loads through a system that was never designed to handle that much watts and amps (it's not really just the voltage after all). The newer chipsets like i925 and i955 for example have a totally redesigned power circuitry that will handle the power sucking Prescotts.

Getting back to the IC7 series, they are the best at Prescott overclocking of any socket 478 mobo and some seem to hold up better than others, but any mobo is only as strong as it's weakest component. Heck, MSI limited vcore on the BIOS updates that supported Prescott. I personally have fried a trusty IC7 at 1.6v with a Prescott. If you read the article in my signature, that was a result of my experience and the desire to help others from suffering the same fate.

There is no magic number like 1.5v is perfectly safe and anything above is guaranteed to go up in smoke. Your mobo survived it's brief tenure at 1.6v without cooling mods. How long would it last? Maybe 1 week? Maybe 1 month? Maybe lots longer? Who knows. But, many others here on the forum killed motherboards in the early Prescott days and they all seemed to happen mainly as we pushed past 1.5v at high clock speeds. So, you need to figure out how much you are willing to risk. Sinking the mosfets and adding case ventilation helps lower the risks. So does limiting voltage. Good luck.

Thanks for the input. Definately opened my eyes to the grim reality of the 875 prescott combo. I definately want to keep this system for longer than a month or until I have to get a 64 bit machine. I think im going to keep with the convention of +- 10% or 1.51v and tweak it as much as I can from there.
Thanks.